| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <1423197085-32270-14-git-send-email-jeffrey.t.kirsher@intel.com>
Date: Thu, 5 Feb 2015 20:31:22 -0800
From: Jeff Kirsher <jeffrey.t.kirsher@...el.com>
To: davem@...emloft.net
Cc: Emil Tantilov <emil.s.tantilov@...el.com>, netdev@...r.kernel.org,
nhorman@...hat.com, sassmann@...hat.com, jogreene@...hat.com,
kernel-team@...com, Alexander Duyck <alexander.h.duyck@...hat.com>,
Jeff Kirsher <jeffrey.t.kirsher@...el.com>
Subject: [net-next v2 13/16] ixgbevf: Add code to check for Tx hang
From: Emil Tantilov <emil.s.tantilov@...el.com>
This patch adds code to allow for Tx hang checking. The idea is to provide
more robust debug info in the event of a transmit unit hang. Similar to the
logic in ixgbe.
CC: Alexander Duyck <alexander.h.duyck@...hat.com>
Signed-off-by: Emil Tantilov <emil.s.tantilov@...el.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@...el.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@...el.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf.h | 21 +++-
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 113 ++++++++++++++++++----
2 files changed, 115 insertions(+), 19 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
index 65cda34..65c2aee 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
@@ -43,6 +43,13 @@
#define BP_EXTENDED_STATS
#endif
+#define IXGBE_MAX_TXD_PWR 14
+#define IXGBE_MAX_DATA_PER_TXD BIT(IXGBE_MAX_TXD_PWR)
+
+/* Tx Descriptors needed, worst case */
+#define TXD_USE_COUNT(S) DIV_ROUND_UP((S), IXGBE_MAX_DATA_PER_TXD)
+#define DESC_NEEDED (MAX_SKB_FRAGS + 4)
+
/* wrapper around a pointer to a socket buffer,
* so a DMA handle can be stored along with the buffer */
struct ixgbevf_tx_buffer {
@@ -85,6 +92,18 @@ struct ixgbevf_rx_queue_stats {
u64 csum_err;
};
+enum ixgbevf_ring_state_t {
+ __IXGBEVF_TX_DETECT_HANG,
+ __IXGBEVF_HANG_CHECK_ARMED,
+};
+
+#define check_for_tx_hang(ring) \
+ test_bit(__IXGBEVF_TX_DETECT_HANG, &(ring)->state)
+#define set_check_for_tx_hang(ring) \
+ set_bit(__IXGBEVF_TX_DETECT_HANG, &(ring)->state)
+#define clear_check_for_tx_hang(ring) \
+ clear_bit(__IXGBEVF_TX_DETECT_HANG, &(ring)->state)
+
struct ixgbevf_ring {
struct ixgbevf_ring *next;
struct net_device *netdev;
@@ -101,7 +120,7 @@ struct ixgbevf_ring {
struct ixgbevf_tx_buffer *tx_buffer_info;
struct ixgbevf_rx_buffer *rx_buffer_info;
};
-
+ unsigned long state;
struct ixgbevf_stats stats;
struct u64_stats_sync syncp;
union {
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index a4b3d66..87f9f86 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -199,14 +199,64 @@ static void ixgbevf_unmap_and_free_tx_resource(struct ixgbevf_ring *tx_ring,
/* tx_buffer must be completely set up in the transmit path */
}
-#define IXGBE_MAX_TXD_PWR 14
-#define IXGBE_MAX_DATA_PER_TXD (1 << IXGBE_MAX_TXD_PWR)
+static u64 ixgbevf_get_tx_completed(struct ixgbevf_ring *ring)
+{
+ return ring->stats.packets;
+}
+
+static u32 ixgbevf_get_tx_pending(struct ixgbevf_ring *ring)
+{
+ struct ixgbevf_adapter *adapter = netdev_priv(ring->netdev);
+ struct ixgbe_hw *hw = &adapter->hw;
+
+ u32 head = IXGBE_READ_REG(hw, IXGBE_VFTDH(ring->reg_idx));
+ u32 tail = IXGBE_READ_REG(hw, IXGBE_VFTDT(ring->reg_idx));
+
+ if (head != tail)
+ return (head < tail) ?
+ tail - head : (tail + ring->count - head);
+
+ return 0;
+}
+
+static inline bool ixgbevf_check_tx_hang(struct ixgbevf_ring *tx_ring)
+{
+ u32 tx_done = ixgbevf_get_tx_completed(tx_ring);
+ u32 tx_done_old = tx_ring->tx_stats.tx_done_old;
+ u32 tx_pending = ixgbevf_get_tx_pending(tx_ring);
+
+ clear_check_for_tx_hang(tx_ring);
-/* Tx Descriptors needed, worst case */
-#define TXD_USE_COUNT(S) DIV_ROUND_UP((S), IXGBE_MAX_DATA_PER_TXD)
-#define DESC_NEEDED (MAX_SKB_FRAGS + 4)
+ /* Check for a hung queue, but be thorough. This verifies
+ * that a transmit has been completed since the previous
+ * check AND there is at least one packet pending. The
+ * ARMED bit is set to indicate a potential hang.
+ */
+ if ((tx_done_old == tx_done) && tx_pending) {
+ /* make sure it is true for two checks in a row */
+ return test_and_set_bit(__IXGBEVF_HANG_CHECK_ARMED,
+ &tx_ring->state);
+ }
+ /* reset the countdown */
+ clear_bit(__IXGBEVF_HANG_CHECK_ARMED, &tx_ring->state);
+
+ /* update completed stats and continue */
+ tx_ring->tx_stats.tx_done_old = tx_done;
+
+ return false;
+}
+
+/**
+ * ixgbevf_tx_timeout - Respond to a Tx Hang
+ * @netdev: network interface device structure
+ **/
+static void ixgbevf_tx_timeout(struct net_device *netdev)
+{
+ struct ixgbevf_adapter *adapter = netdev_priv(netdev);
-static void ixgbevf_tx_timeout(struct net_device *netdev);
+ /* Do the reset outside of interrupt context */
+ schedule_work(&adapter->reset_task);
+}
/**
* ixgbevf_clean_tx_irq - Reclaim resources after transmit completes
@@ -311,6 +361,37 @@ static bool ixgbevf_clean_tx_irq(struct ixgbevf_q_vector *q_vector,
q_vector->tx.total_bytes += total_bytes;
q_vector->tx.total_packets += total_packets;
+ if (check_for_tx_hang(tx_ring) && ixgbevf_check_tx_hang(tx_ring)) {
+ struct ixgbe_hw *hw = &adapter->hw;
+ union ixgbe_adv_tx_desc *eop_desc;
+
+ eop_desc = tx_ring->tx_buffer_info[i].next_to_watch;
+
+ pr_err("Detected Tx Unit Hang\n"
+ " Tx Queue <%d>\n"
+ " TDH, TDT <%x>, <%x>\n"
+ " next_to_use <%x>\n"
+ " next_to_clean <%x>\n"
+ "tx_buffer_info[next_to_clean]\n"
+ " next_to_watch <%p>\n"
+ " eop_desc->wb.status <%x>\n"
+ " time_stamp <%lx>\n"
+ " jiffies <%lx>\n",
+ tx_ring->queue_index,
+ IXGBE_READ_REG(hw, IXGBE_VFTDH(tx_ring->reg_idx)),
+ IXGBE_READ_REG(hw, IXGBE_VFTDT(tx_ring->reg_idx)),
+ tx_ring->next_to_use, i,
+ eop_desc, (eop_desc ? eop_desc->wb.status : 0),
+ tx_ring->tx_buffer_info[i].time_stamp, jiffies);
+
+ netif_stop_subqueue(tx_ring->netdev, tx_ring->queue_index);
+
+ /* schedule immediate reset if we believe we hung */
+ schedule_work(&adapter->reset_task);
+
+ return true;
+ }
+
#define TX_WAKE_THRESHOLD (DESC_NEEDED * 2)
if (unlikely(total_packets && netif_carrier_ok(tx_ring->netdev) &&
(ixgbevf_desc_unused(tx_ring) >= TX_WAKE_THRESHOLD))) {
@@ -1479,6 +1560,8 @@ static void ixgbevf_configure_tx_ring(struct ixgbevf_adapter *adapter,
txdctl |= (1 << 8) | /* HTHRESH = 1 */
32; /* PTHRESH = 32 */
+ clear_bit(__IXGBEVF_HANG_CHECK_ARMED, &ring->state);
+
IXGBE_WRITE_REG(hw, IXGBE_VFTXDCTL(reg_idx), txdctl);
/* poll to verify queue is enabled */
@@ -2643,6 +2726,12 @@ static void ixgbevf_watchdog(unsigned long data)
if (test_bit(__IXGBEVF_DOWN, &adapter->state))
goto watchdog_short_circuit;
+ /* Force detection of hung controller */
+ if (netif_carrier_ok(adapter->netdev)) {
+ for (i = 0; i < adapter->num_tx_queues; i++)
+ set_check_for_tx_hang(adapter->tx_ring[i]);
+ }
+
/* get one bit for every active tx/rx interrupt vector */
for (i = 0; i < adapter->num_msix_vectors - NON_Q_VECTORS; i++) {
struct ixgbevf_q_vector *qv = adapter->q_vector[i];
@@ -2656,18 +2745,6 @@ watchdog_short_circuit:
schedule_work(&adapter->watchdog_task);
}
-/**
- * ixgbevf_tx_timeout - Respond to a Tx Hang
- * @netdev: network interface device structure
- **/
-static void ixgbevf_tx_timeout(struct net_device *netdev)
-{
- struct ixgbevf_adapter *adapter = netdev_priv(netdev);
-
- /* Do the reset outside of interrupt context */
- schedule_work(&adapter->reset_task);
-}
-
static void ixgbevf_reset_task(struct work_struct *work)
{
struct ixgbevf_adapter *adapter;
--
1.9.3
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists